DMX project - to ground or not to ground?

[cedd]

Hi all

I've built some kabuki-esque drop devices for a couple of shows coming up. They'll be dropping lightweight cloth "trees" that will drop and unravel for the scene and then the whole thing will drop at the end. Everything very lightweight before people start complaining about DMX being used to drop things - don't worry. I can foresee them being useful in the future for panto and other shows where lightweight things like stuffed toys, letters, fabric etc. need dropping on cue.

I'm using an arduino nano and custom PCB I've designed and had etched (MAX485's, opto isolated, dip switch addressing etc). I've got DMX in and out ports (passive linked through) which connect all 3 pins to each other. My query is if I should connect the DMX ground to my pcb ground or not. As a differential signal I don't THINK I need to for them to work. As I'm currently testing on a bench with a common ground though it's difficult to know as it would just work anyway. Concerns are ground loops etc. if DMX ground and PSU ground end up at different potentials. Power supply to my units is just a 2.1mm DC socket (I'm running DC up to the grid for them from a common supply, via speakon cables with an adapter box to break out to 2.1mm at each fixture) so they could feasibly be used with a wall wart or similar in the future. So "ground" may not necessarily be ground anyway.

 

For those interested, they use a payload/tow release device designed for model gliders etc. Really nice aluminium unit bought from Hobbyking. Then just a servo to drive them. Each unit can operate 2 servos (one for release to unravel, one to drop at the end) and has a switchable (by bit 10 on the address dip switch) safety channel mode whereby the next channel up from the "drop" channel has to be at 50-60% before it'll drop, to prevent an "all @ full" releasing everything. The whole thing is built in to a die cast enclosure with the front panel replaced with a 3d printed plate for all the connections and hole for the dip switch.

Quite pleased to now have a reliable and working way of receiving DMX in to an Arduino. All sorts of opportunities open up now I can do it affordably.

 

Thanks for your time

[martinw]

Hello

 

You should connect pin 1 of the DMX connector to 0V ref plane on the isolated side of your opto-isolation barrier (I'm assuming that this also has an isolated PSU which is separate from the main PSU system of your device. Connecting pin 1 ensures that the data signals are within the common-mode range on the MAX485, but having an isolation barrier via the opto isolators and isolated PSU prevents any kind of ground loops. See page 11 of the DMX spec "Isolated Receiver" - this is the prefered way to build the receiver and results in the most robust system.

 

Martin

[DrV]

Yes you should connect it. RS485 is not a true differential signal - the legs go between 0 and about +5. They don't go -ve. As stated above you need to keep the common mode voltage in that range for the receivers to work reliably

[maeterlinck]

Ideally as suggested you would have a fully isolated side (part exposed to the DMX bus), so opto isolator for the signal and power isolation for the system power. ICs that will do all of this in one device are available but they are expensive.

 

It is possible to 'get away' without doing so.

 

Terminating the DMX ground to the device ground will help prevent some faults. Equally leaving it floating will prevent others (in a non isolated system).

It can also be tied to ground via a 10K resistor.

[cedd]

Thanks folks. You can tell I'm an audio guy - I've always thought of RS485 as essentially a differential balanced line, so it's interesting to hear it's not quite as true as I'd thought.

Right, looks like grounded (as I'm isolated) is the way forward. Makes perfect sense when I think about it - that's kind of the whole point of being isolated isn't it!

[timsabre]

Well it is a differential signal, but the receiver chip can only handle a common mode voltage of max 12V, above which everything stops working. So you need to use the ground reference to keep the +5V/0V differential signal within a 12V range of the receiver's ground.

 

Presumably the same applies to audio balanced line using an op-amp to debalance the signal - it'd have to be within the power rails of the op amp. That's usually more than 12V though.

[DrV]

Well it is a differential signal, but the receiver chip can only handle a common mode voltage of max 12V, above which everything stops working. So you need to use the ground reference to keep the +5V/0V differential signal within a 12V range of the receiver's ground.

 

Presumably the same applies to audio balanced line using an op-amp to debalance the signal - it'd have to be within the power rails of the op amp. That's usually more than 12V though.

Exactly but for an op-amp the range usually also extends below zero. Typical supply rails in a mixer being +/- 15V.

[timsabre]

The normal DMX chip, 75176 will also go negative - it will work over the range -7V to +12V.

I just meant I'd never thought of it for audio...

 

[sunray]

The normal DMX chip, 75176 will also go negative - it will work over the range -7V to +12V.

I just meant I'd never thought of it for audio...

This is why I so much prefer audio I/p's & o/p's to have isolating transformers... twisted non screened pairs galore, run directly into telecom cables etc etc.

 

RS232 runs over +/-12 and with care RS485 can be interfaced (almost) directly, so from that I'd expect RS485 chips to be capable of the same levels.